Rivian says its battery pack has the highest volumetric energy density in the world, thanks to its cooling strategy

Rivian has been generating a lot of headlines since the November 2018 reveal of its prototype pickup and SUV. The company has scored substantial investments, including $700 million from a group of investors led by Amazon and another $500 million from Ford. That cash was accompanied by a history-making order of 100,000 electric delivery vans from Amazon and plans to partner with Ford to build new EVs (including a rumored electric Lincoln).

Charged recently chatted with Rivian’s VP of Propulsion Richard Farquhar to learn more about the design and development of the company’s unique electric platform (you can read the whole story in our upcoming issue).

Farquhar described how Rivian engineers were able to achieve unprecedented energy density.

Richard Farquhar: We have what we believe is currently the highest volumetric energy density modules and pack in the world. It’s about 20-25% greater energy density in watt-hours per liter compared to anything on the market today.

What has allowed us to achieve this is the construction of the module. Each module has two layers, each containing 21700 type cylindrical cells. So, we’ve got 15 kWh of energy in each module. There are 9 of those modules in our standard pack, and we have 12 of those in what we call our premium pack of 180 kWh, which gets us over 400 miles of range.

At the heart of the module is a cooling plate between the upper and lower cell layers. That allows us to control the cooling of those cells in their most efficient medium, which is cooling axially. We pull the heat out of the cell through its center, which is the most efficient way to do it, as opposed to radially. It allows us to pack those cells really close together. This allows us to get the highest volumetric energy density available today.

Rivian’s Battery Pack

Because it cools the cells axially instead of radially, Rivian was able to design its modules with only air between the cells. There are no liquid cooling channels (or other thermally conductive materials) weaving between the cells like the ones Tesla and others have used.

Rivian is set to launch production at the end of next year. Farquhar told Charged the team is currently in the validation, reliability and durability testing phase with “tens” of production-intent vehicles.

  • Adrian Tatum

    Someone help me, please. With its Model S Long Range, Tesla gets 373 miles from a 100 kWh pack— and that’s from the old 18650 cells. They could increase capacity just a smidge at be at 400. Furthermore, changing to the 2170 cells used in the Model 3 would blow the roof off the suckers.

    • Vincent Wolf

      It’s to bad Mini doesn’t have an agreement with Tesa for it’s Mini which it’s pathetic 110 mile range on it’s new 2020 Mini EV is a joke. That’s the same as the 2017 Nissan Leaf.

      • TheIceismelting!

        Why would Tesla allow anyone to charge at its stations except when forced to as in the EU? tesla wants to make a profit, not expand EV adoption.

        • Robert

          I don’t know if I want to be in a Tesla den with a non Tesla. A couple of times I pulled into a Harley nest on my BMW R 1200RS…did not feel inviting,

    • scottf200

      1000 lbs heavier wind pushing truck vs lighter slippery sedan. Way different tires (friction stills watts). You are not even close to a fair comparison.

  • http://www.linkedin.com/in/ablelawrence Able Lawrence

    Liquid or gel cooling is more efficient.
    Tesla has also patented phase changing gels that would then buffer heat and maintain constant temperature.

  • morrisg

    Rivian has adopted an interesting battery strategy. They are making full 96 cell series stacks in each module. There are a number of cells in parallel at each step in the series stack. So each module has the full 400 volts and can be easily paralleled with other modules to increase kw-hr capacity. So in theory a customer could buy a short range version, discover they really need more range (change of job?) and purchase additional battery modules later. Of course, that might upset their marketing price steps between models within the range, but there’s no electrical reason it couldn’t be done.

    As far as the cooling comment by Richard Farquhar, I’m not sure I believe axial cooling is more efficient than radial cooling to a significant degree, maybe a little but not a huge difference. The backing electrodes in the jelly roll are usually aluminum and copper which readily conduct heat, but these electrodes must be electrically isolated from the metal case of the 21700 cell. So how does the heat transfer from the electrodes to the bottom of the case? Does this depend on liquid electrolyte? And is the 21mm diameter of the case bottom large enough to conduct all of the internally generated heat without appreciable temperature rise?

    Another question is the heat capacity of the cold plate, which will depend on the dimensions and flow rate of the coolant through it. I hope they have overdesigned it so it can cover even the corner cases of supercharging in Arizona summer heat.

    I was impressed with their use of 21700 cells and adopting ultrasonic wire bonding technique to make both + and – connections at one end of each cell thereby allowing full contact between the bottom of the cell and the cold plate.

    All of this came from examining photos of their battery lab published here on Charged EVs about a year ago. Yes, I’m an electrical engineer who has done simple small lithium ion battery designs but I’m retired now. Thanks for a good article, looking forward to the full interview.

    • http://www.aeva.asn.au Chris Jones

      Correct – the thermal conductivity is two orders of magnitude higher in the axial dimension, however there is a big thermal resistance between the jellyroll and the end of the cell. Moreover, it’s better through the negative end than it is through the positive end. I wasn’t sure if Rivian were only terminating one end of the cell or not, but it would make sense to do that, as the base of the cell could be in good contact with the cooling plate. There is still an electrically isolating, thermally conductive adhesive between the base of the cell and the cooling plate.

      • morrisg

        I think the axial direction heat conductivity is higher simply due to the metal electrodes being continuous in that direction compared to radially where the heat must go through the cathode chemistry, separator, anode chemistry, electrode, etc as the heat travels through the jelly roll. However, if we think of the metal electrodes and case having so much higher heat conductivity compared to the chemistry, then we can approximate their temperature as a constant across the metal surface. We know it’s not actually constant, but due to the high heat conductivity it could be small, maybe one or two degrees C? So if this is true (depends on actual conductivity values and amount of heat conducted) we could say there’s only a small difference between radial and axial coldplate cooling efficiency due to the case being almost an isobar of constant temperature.

        Another aspect of radial vs axial thermal performance is the area of the case that is in contact with the cooling plates. This will affect the temperature rise for the total amount of heat transferred. So if the radial contact area between the cold plate and the side of the case is 2x to 3x the area of the axial contact between the bottom of the case and the cold plate, that could reduce thermal resistance for the radial case.

        Of course, all of this also depends on the amount of heat needed to be conducted away from the cell, the actual thermal resistances involved and therefore the temperature rise experienced by the cell chemistry.

        Gee, I’m enjoying an actual engineering discussion of the various cases on a comment board! Thank you!!!

        • Erkko

          The casing is typically made of stainless steel or nickel plated A3 steel, which have poor thermal conductivity (5-15% of copper). The material is very thin, so the heat does not spread evenly and instantly.

  • EddyKilowatt

    Putting radial air gaps between cells allows Tesla (and others?) to thread their coolant conduit in between them, but it also serves the useful purpose of decreasing thermal coupling between cells in the event of a thermal event in one cell. Or in other words, cell spacing assists with preventing or at least reducing propagation in the event of fire or thermal runaway. I wonder how Rivian is handling this rare, but obviously potentially catastrophic, scenario in their pack?

    • scottf200

      Whenever there is a ‘thermal event’ in a Tesla it is not a controlled event. I’m a Tesla fan and owner but watch the videos of several scenarios. It is not minor.

      • EddyKilowatt

        Well yeah, things can and do go wrong in a big way… but we don’t know how often one cell in a pack has an internal short, gets hot and maybe even vents, blows its bond-wire fuse, but *doesn’t spread to adjacent cells* and then to the whole car. Safety systems that work don’t wind up on the evening news and YouTube. Not that a 1/8″ air gap is a bulletproof safety system, mind you… but I bet there will be differences in propagation behavior versus zero gap between cells.

        • Erkko

          The NCA chemistry used by Tesla doesn’t need an internal short. It goes into thermal runaway when the temperature of the cell rises above 160 C. Packing the cells with insulating foam and phase-changing materials which absorb heat is necessary to control and slow down the chain reaction. Otherwise the Tesla batteries would have a much higher probability of lighting up from the slightest damage and burning down the car and passengers before any rescuers can get to the crash.

          Especially the Model 3 battery has a new heat-expanding foam which insulates and pushes out the burning cells so the fire wouldn’t spread so easily. Packing the battery tighter with bigger cylinders and having no insulation material between makes a battery into a fire bomb.

  • homeofthebrave

    Heating in electrical systems is lost energy. When efficiency is high (as it is in supercapacitors) the I2R losses are very low, little heat generated. Thus having an ‘advanced’ cooling system to rid a system of heat is an indication of low efficiency (wasted energy). Also, is no one looking at basic vehicle energy physics? That is, a vehicle having 2 x the weight of a standard vehicle has to consume 2x the friction energy, 2x the momentum energy (albeit partially recoverable), 2x the elevation energy (also partially recoverable). Thus how is a 5,000 lb Tesla saving energy, over my 1900 lb 1982 Honda?

    • morrisg

      This is easy to estimate: My Tesla Model 3 weighs 3,805 lbs which is about 2x what your 1900 lb 1982 Honda weighs. My car uses less than 250 watt-hrs of electricity per mile of driving, averaged over about 15,000 miles. A gallon of gas has about 36.6 kw-hrs of energy in it. So my Tesla Model 3 goes 146.4 miles on the electrical equivalent of one gallon of gas. How far does your Honda go on one gallon of gas? Using the lookup tool on fueleconomydotgov, it only goes back to 1984 but the 1984 Honda Civic is listed as 29mpg under new calculation rules and 39mpg under old calculation rules. Thus, my Tesla Model 3 Long Range Single Motor goes 146.4/39= 3.754 times further than your Honda on the same amount of energy!

      • Erkko

        Don’t forget that batteries have embedded energy costs (ESOEI). A lithium battery costs about 10% of its lifetime energy storage capacity to manufacture. The lifetime capacity is compared to the potential of storage, which is greater than what you end up using. E.g. a battery that holds 100 kWh and lasts for 2,200 charge cycles costs about 22,000 kWh to make – or about enough energy to run an average US household for a year.This is still only the nominal cost, because the actual battery has extra capacity built-in to account for normal capacity drop, and it doesn’t account for the other features of the battery such as the BMS, cooling system, and armouring.

        Now, if you manage 15,000 miles a year for 10 years when the battery meets its shelf-life (lithium batteries wear out even when not in use) and you have to replace it; at 250 Wh per mile you’ve only managed to put 37,500 kWh out of a possible 220,000 kWh through the battery, and this number compares to the manufacturing cost of the battery. You are actually spending 159% the energy to drive those miles.

        Then you have to take into account the efficiency of the electric grid and the charging system, which isn’t 100%. The average grid loss is about 6-7% and the charger is 90+% for approx. 85% efficiency. Imagine if you lost 15% of your gasoline at the pump!

        Then you need to account for the average efficiency of the generator stations – because you can’t actually choose which electrons you take from the grid. If a wind turbine is turning and you plug in your charger, the wind won’t magically pick up to make more power for you. What actually happens is, a gas turbine somewhere turns up a little. Give it a generous 50% and re-calculate. The energy used becomes 88235 + 22,000 kWh = 110,235 kWh which is 2.94 times the initial estimate. This means your estimated energy advantage of 3.754 times shrinks to 1.27 times.

        Which surprisingly means, your Tesla is actually consuming very close to the same amount of energy in total as a 1982 Honda – because the extra cost is hidden in the infrastructure and the embedded energy to make the battery. The energy cost to transport tanker trucks full of fuel is less than a thousandth part of the energy contained within – the electric grid is actually a relatively inefficient means to process and transmit energy, and batteries need minerals and materials, water, electricity, fuels…literally tons for every car.

        This is also the reason why other EV manufacturers are using much smaller batteries. When the battery is small, the embedded energy loss is small. There’s no real sense in designing a huge battery because the extra capacity largely goes to waste – you can’t physically drive it enough to exhaust the miles before the battery dies of old age.

  • jstack6

    What is the Wall hour per liter capacity of the cells. I don’t see that.
    Also what is the cost per kWh?

  • Vincent Wolf

    It sure will be nice when the entire industry standardizes batteries and modules and anyone can buy additional modules to upgrade their range funds permitting. Competition is great to see but the world really needs standardization much like Detroit cars all look the same.